ICU and You · Myths and Misconceptions · Number 1
The school-age child — where each belief came from, and what the evidence actually says
The aphorisms on this site take a single saying and turn it over — they are about wisdom, they are contestable, and the honest verdict is often maybe.
This series is about facts. Several per edition, each one checkable, each one something a competent clinician currently believes. But every entry runs the same four steps — the claim, where it came from, what the evidence shows, and a verdict — because the origin is the most interesting part.
Every durable myth was once a reasonable inference. Knowing why sensible people believed a thing is the only part of this that transfers, because it is how you recognise the next one while it is still forming. The verdicts here are graded rather than gleeful, except where the matter really is settled.
All five of these are things I was taught, taught to others, or both.
“The cricoid ring is the narrowest part of a child's airway — which is why you never use a cuffed tube under the age of eight.”
In 1897 a French anatomist named Bayeux made plaster casts and sections of the larynges of fifteen children who had died of diphtheria, aged four months to fourteen years. In all fifteen, the cricoid circumference was smaller than the glottis. In 1951 Eckenhoff reviewed that work and carried it into the English-language anaesthetic literature — adding, to his credit, the caution that measurements taken from the dead “may not be completely applicable to the living”.
Fifteen dead children, one nineteenth-century observer, and an explicit warning from the man usually blamed for it. That is the entire evidential base for a doctrine that governed paediatric airway practice for sixty years. The “under eight” cut-off has no identifiable primary source at all — it appears to be a teaching heuristic that accreted.
The glottis is the narrowest aperture, but it is pliable — under paralysis the cords abduct and offer a passing tube almost no resistance. The cricoid remains the narrowest non-distensible ring, which is why it is still where mucosal injury happens.
And the ellipse matters practically: an uncuffed tube that leaks audibly can still be pressing hard on the lateral tracheal wall. The audible leak was never the reassurance we treated it as.
The cricoid is not the narrowest part of the airway; the glottis is. It is still the narrowest rigid part, so the injury concern was pointing at the right structure for the wrong reason.
But do not over-correct. ANZCOR's position is that cuffed and uncuffed tubes are both acceptable in infants and children, with attention to cuff pressure. The European guidance says use cuffed. So “cuffed is mandatory” is, in Australia, as unsupported as “uncuffed is mandatory” — what is dead is the idea that a cuff is dangerous. Sizing: uncuffed age/4 + 4, cuffed age/4 + 3.5, and use a manometer, because palpating the pilot balloon is not a measurement.
“Give fluid too fast, or with the wrong sodium, and you will cause cerebral oedema in diabetic ketoacidosis.”
A 1988 series of 42 children, with no comparison group, in which a faster fluid rate was associated with a shorter time to herniation. Around it grew a mechanistic story that was beautiful and persuasive: idiogenic osmoles accumulate in brain cells, plasma osmolality falls as you treat, water follows the gradient inward, the brain swells.
It explained a terrifying complication, it identified a culprit, and — the part that really entrenches a belief — it gave clinicians something to do about it. A 2001 case-control study found no association with fluid volume at all, and reported instead that low arterial carbon dioxide, high urea and bicarbonate therapy carried the risk. It changed rather little.
A randomised trial across thirteen American emergency departments enrolled 1,389 episodes of DKA in 1,255 children, in a two-by-two design: fast rehydration against slow, crossed with half-normal against normal saline.
Decline in mental status occurred in 3.5 per cent of episodes and clinically apparent brain injury in 0.9 per cent, with no significant difference by rate or by sodium concentration — not on mental status during treatment, not on short-term memory, and not on memory or IQ months later. The only faint signal pointed the other way: a non-significant trend towards faster cognitive recovery in the rapidly rehydrated children.
The mechanism has moved with it. Current thinking is cerebral hypoperfusion during the illness, reperfusion during treatment, on a background of inflammation — with the injury correlating with how dehydrated and how hyperventilating the child was on arrival, rather than with anything that happens to osmolality afterwards. Much of the damage is under way before the first bag is hung.
That trial counted twelve clinically apparent brain injuries in total. It robustly excludes an effect on mental status and neurocognition. It cannot exclude a modest effect on rare catastrophic herniation, and nobody should read “no difference across 1,389 episodes” as “no difference in the thing that kills children”.
Nor did it test extremes: the slow arm assumed a 5 per cent deficit over 48 hours and the fast arm 10 per cent with two boluses. The two arms also differed in assumed deficit and number of boluses, so this compares two protocols rather than isolating rate. Nothing in it licenses unrestrained fluid.
Within the range any reasonable protocol uses, neither the speed of rehydration nor the tonicity of the fluid drives cerebral injury. International guidance now replaces the deficit over 24 to 48 hours with anything from half-normal to normal saline, and explicitly states that deliberately driving the sodium up is no longer considered necessary. Australian practice sits at the conservative end of that range — 48 hours, isotonic — which is a choice within the guideline, not a disagreement with it.
Cerebral injury in DKA is real, is uncommon, and kills or maims a substantial share of the children it affects. It is driven by how sick the child was when they arrived, not by your drip rate. Vigilance is still owed. Guilt is not.
“Estimate the burn with the rule of nines — and the patient's palm is one per cent.”
The rule of nines is a genuinely good adult heuristic, and heuristics travel. Lund and Browder published the age-corrected chart in 1944, modifying Berkow's earlier adult proportions, and their innovation was precisely the thing the nines lacks: a correction factor for children. Eighty years later it is still the standard, and it is still routinely skipped.
A child is a different shape. The head is around 19 per cent of body surface at birth against 7 per cent in an adult; the thighs run the other way, roughly 11 per cent combined at birth against 19 in the adult. The rule of nines gives the head 9. Apply it to a young child and you understate the head and overstate the legs — which are the two regions scalds most often involve.
When adult proportions become valid is genuinely disputed. NSW guidance has you subtract one per cent from the head and add half a per cent to each leg for every year of life until age nine. The Lund and Browder bands themselves do not reach adult values until fifteen. Either way, nobody in the school-age range has adult proportions.
And the palm. The correct rule is the palm plus the fingers — the whole palmar surface — at approximately one per cent. The palm alone is about half that. Teaching “the palm is one per cent” roughly doubles the true figure, and it does so precisely in the patchy burns where the palmar method is used.
A Sydney audit of 698 patients transferred to adult burns units found overestimation exceeded underestimation by more than three to one, and that only 30 per cent of referring estimates matched the burns unit's. A systematic review of 28 studies and 6,461 patients found overestimation in at least half of cases in most comparative studies — one paediatric cohort overestimated by 100 per cent — and, in one series, 55 per cent of patients received more than 125 per cent of the recommended fluid volume.
Overestimation means over-resuscitation: oedema, compartment syndrome, prolonged ventilation, and children retrieved who did not need to be. Underestimation, which rises with larger burns, delays resuscitation in the children least able to tolerate delay.
Use an age-specific chart in any child. The adult rule of nines is simply wrong in a school-age child, and the palmar rule means palm and fingers.
The harder second sentence: the practical failure here is not ignorance of the chart. It is confidence without it.
“Cool them after cardiac arrest.”
Adult trials, animal work and a thoroughly plausible mechanism — reduced metabolic demand in an injured brain. For about fifteen years therapeutic hypothermia went from novel to standard of care, and it was taught as such. This one is a useful corrective to the idea that myths are always old.
Two randomised trials, out-of-hospital and in-hospital arrest, 295 and 329 children. Both compared 33°C against 36.8°C. Both were null — 20 per cent against 12 per cent for good functional survival in the first, not significant; 36 against 39 per cent in the second, which stopped for futility.
In both trials the comparator arm was not usual care. It was actively maintained at 36.8°C, using the same cooling device, for a full 120 hours. Nobody was allowed to drift febrile.
So what the trials show is that hypothermia is not superior to controlled normothermia — not that temperature does not matter. A trainee who takes “cooling doesn't work” from these trials has inverted them, and will allow a post-arrest child to sit at 38°C.
Guidelines have landed exactly there. Continuous central temperature monitoring in the comatose child after arrest is a strong recommendation on good evidence; avoiding temperatures above 37.5°C is a strong recommendation on weaker, observational evidence; and either a period at 32–34°C followed by normothermia, or normothermia alone, is reasonable. ANZCOR gives you a ceiling rather than a range — actively control to 37.5°C or below.
Worth being honest about the weak link: the evidence that fever after arrest harms children is observational, not randomised. We are confident enough to mandate it, and that confidence does not come from a trial. Saying so is better teaching than pretending otherwise.
Don't cool. Control. Letting a comatose post-arrest child run above 37.5°C is a departure from every current guideline, and it is the far commoner error now that everyone has heard the trials were negative.
“Nil by mouth from midnight.”
Mendelson's 1946 description of aspiration in obstetric anaesthesia, generalised to everybody, and then made operationally convenient. Midnight is not a physiological time. It is an administrative one — a list order is uncertain, so the safe institutional move is to fast everybody from the earliest conceivable start.
A European consensus in 2018 recommended clear fluids up to one hour before elective anaesthesia in children. Australian guidance has followed: ANZCA's current professional document allows clear fluids at 3 mL/kg/hr up to one hour beforehand, with solids at six. Some Australian units have gone further and run “sip till send”.
And the safety question has now been answered at a scale that is hard to argue with. An international prospective cohort of 306,900 anaesthetic procedures across 31 centres compared sip-till-send, one hour and two hours. Confirmed pulmonary aspiration ran at 1.18, 0.96 and 1.83 per 10,000 respectively. Both liberal regimens were non-inferior — and the point estimates were numerically lowest in the liberal arms.
Now the number that should bother you. An Australian audit of a simple “apple juice on arrival” protocol found that before the change, children were fasting from clear fluids for a mean of 9.9 hours. Afterwards, 3.5. Preoperative blood glucose rose, severe hypoglycaemia was eliminated, nobody aspirated, and the complaints from parents stopped.
Breast milk is three hours, not four, in the current Australian document — a number many of us were taught differently. Irrelevant to a ten-year-old, and worth knowing anyway.
The one-hour rule is for healthy children having elective anaesthesia. Reflux, gastroparesis, achalasia, severe cerebral palsy, enteropathy and renal failure are genuine exceptions, and none of this applies to emergency surgery.
Nobody writes “fast from midnight” in a guideline any more. The guidelines changed and the wards did not, and children are still going ten hours without a drink to satisfy a list order that could have been communicated.
Which makes this the one on the list you can personally fix this week, and the only one where the fix is a conversation rather than a paper.
What were you taught that turned out not to be true?
Everybody has one. A rule you followed for years, a number you repeated confidently, a contraindication that turned out to be somebody's habit. I am as interested in when you found out and how it felt as in the fact itself.
I am also interested in the opposite case — something dismissed as an old wives' tale in your training that turned out to be right. Those are rarer and more instructive.
Send them and I will work the best into a future edition, named or anonymous exactly as you prefer.
The button opens a reply straight to me. If your device blocks it, write to icuandyou@icloud.com instead.